The growing number of studies have concluded there is an evident correlation between rises in ambient air pollutants' concentration and adverse birth outcomes, nevertheless there are still inconsistent reports on the degree of an influence. The studied published meta-analyses mostly agree on the high heterogeneity between original studies but fail to substantially profile its sources.
The aim of our study is to gather newer studies as a part of a systematic review and include them in our existing database of the research of Klepac et al 2018. We are conducting a meta-analysis in order to compare merged effects of exposure assessment method, type of pollutant and window of exposure. Furthermore, we are assessing each study based on its properties and covariates that might variate between studies. Based on these results, we are intending to develop meta-regression models for each pair of pollutant – adverse and birth outcome that would help to explain the heterogeneity of discovered effect estimations between studies.
Systematic review is conducted for the period of 7.11.2017 – 21.2.2020 on Pubmed, a search engine for the Medline, the bibliographical database. Included studies must be original research work, that study the relation of mothers’ exposure of pollutants of particulate matter with the diameter less than 10 µm (PM10), less than 2,5 µm (PM2.5) and gasses: nitrogen dioxide (NO2), ozone (O3) or carbon oxide (CO) for the adverse birth outcomes of preterm birth (PTB) or low birth weight (LBW). The former could also be studied as an absolute difference of the birth weight (BW). The quality of the studies' methodology is assessed with the Newcastle-Ottawa Scale manual. Meta-analysis is conducted first in Review Manager 5.4.1 and then again in STATA 16.0. Subgroup analysis and meta-regression model development are also performed in the latter. Subgroup analysis includes data on methodology, geographical area, population, inclusion criteria, covariates for regression models’ effect adjustment, average concentrations and NOS-based quality score. Models are developed in a way that they explain as much heterogeneity as possible while staying statistical significant as entities.
Meta-analysis included 121 studies. For the outcomes of PTB and LBW the results are presented as unified odds ratios (OR) and difference in birth weight in grams is presented as factor β. We find statistically significant correlations between rises in concentration of PM10, PM2.5, NO2, O3 for 10µg/m3 and CO for 100µg/m3 in the entire pregnancy in studies which applied MS exposure assessment method. Results for those that applied MA method were not as consistent, possibly because of the lower number of studies. Amongst pollutants, the PM2.5 has the greatest effect on the chosen outcomes (results from STATA 16.0; entire pregnancy): 1,094 with 95% CI [1,032; 1,159] for PTB, 1,040 with 95% CI [1,016; 1,064] for LBW, -17,11g with 95% CI [-23,04; -11,18] for BW. Results for studies that apply the method MA provide similar results. Among the trimesters we find the rises in concentration of PM10 and PM2.5 in the second trimester to be the most critical in the PTB.
We can conclude that ambient air pollution, especially PM2.5 is a considerable factor in the preterm birth and (low) birth weight. Even though rises PM10 and NO2 are shown to have marginal effects in comparison to PM2.5, they may also be a factor in the cumulative burden of the influence of air pollution on the studied outcomes. The results for the rises in concentration of O3 and CO are shown to have a significant influence on certain outcomes, however they are still too variable to enable us to declare a definitive conclusion. Our meta-regression models have proven to be possibly applicable only for the outcome of birth weight, with comparison for a global region and adjustment for a socioeconomic status of the mother or the household (SES) as the key covariates.